CHEVRON AND DEFERENCE IN STATE ADMINISTRATIVE LAWfordhamlawreview.org/wp-content/uploads/assets/pdfs/Vol_83/Saiger... · 555 CHEVRON AND DEFERENCE IN STATE ADMINISTRATIVE LAW Aaron

555

CHEVRON AND DEFERENCE

IN STATE ADMINISTRATIVE LAW

Aaron Saiger*

INTRODUCTION

Chevron1 solves a multidimensional conundrum. Whether and how courts should defer to agencies, when those agencies interpret the statutes that grant them authority to make law, is a complex constitutional puzzle. But any solution to that puzzle also reverberates as a matter of political theory, of jurisprudence, and of public administration. Chevron‘s principle, that courts should defer to any reasonable interpretation by an agency of ambiguities in a statute that empowers it to act with the force of law,2 is pragmatic, responsive, and elegant with respect to bedeviling issues in each of these domains.

As a matter of constitutional law, Chevron provides an account of congressional delegation of executive power, explicit but especially implicit, that is genuinely plausible and that is consistent with a genuine doctrine of separation of powers.3 As a matter of political theory, Chevron offers a workable equilibrium between the elitism of technocrats and the imperative for democratically accountable politics.4 As a matter of jurisprudence, its formula gives serious content to the principle that courts must say what the law is, while still assuring to agencies genuine deference, rather than a faux psuedo-deference by which courts defer to agencies when and if they agree with them.5

From the perspective of statutory law, Chevron famously does not cite the federal Administrative Procedure Act6 (APA).7 But, in its structure,

* Professor of Law, Fordham University School of Law. This Essay is part of a larger symposium entitled Chevron at 30: Looking Back and Looking Forward. For an overview of the symposium, see Peter M. Shane & Christopher J. Walker, Foreword: Chevron at 30: Looking Back and Looking Forward, 83 FORDHAM L. REV. 475 (2014). I thank the organizers and participants in the symposium for their bracing comments and responses; my Fordham colleagues Nestor Davidson and Ethan Leib for additional, valuable comments; the Fordham Law School for financial support; and Laurence Abraham and Todd Melnick of the Fordham Law Library for superlative reference assistance. 1. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837 (1984). 2. Id. at 866. 3. See Emily Hammond Meazell, Presidential Control, Expertise, and the Deference Dilemma, 61 DUKE L.J. 1763, 1774–75 (2012). 4. See Frederick Liu, Chevron as a Doctrine of Hard Cases, 285 ADMIN. L.J. 285, 315–16 (2014). 5. See id. at 321–23. 6. 5 U.S.C. § 706 (2012).

556 FORDHAM LAW REVIEW [Vol. 83

Chevron honors the APA‘s categories, or, perhaps more precisely, drinks from the same sources of inspiration. Both Chevron and the federal APA emphasize reasonableness as the sine qua non of lawful agency action and the basis for judicial review.

Finally, and not incidentally because of these features, Chevron succeeds as a matter of public administration. It provides agencies with genuine spheres of action where they can deploy expertise and expect deference.8 It both creates and limits space for politics in agency decisions. It negotiates between the Scylla of congressional micromanagement of statutory implementation and the Charybdis of ad hoc, unpredictable judicial review.

Of course, not one of these claims for the virtue of Chevron is uncontested. Much of the copious literature on the case is devoted to interrogating, and in some cases rejecting, the claims I mention. Perhaps Chevron does not hold up with respect to a truly serious account of congressional intent.9 Maybe it offers no genuinely coherent articulation of the separation of powers.10 It arguably misapprehends executive power11 and the APA.12 It is ―thin‖ and arguably naive in its treatment of agency expertise.13 It may not provide judges with a workable or enforceable delineation of when to review de novo and when to defer.14 One can think that Chevron constrains agencies too much, or not enough.15 But one of the key features of Chevron is its plausible claim to address the question of deference with respect to all of these issues at once.

This Essay suggests some justifications for the somewhat surprising fact that Chevron, with its overwhelming impact upon federal administrative

7. See Thomas J. Miles & Cass R. Sunstein, Do Judges Make Regulatory Policy? An Empirical Investigation of Chevron, 173 U. CHI. L. REV. 823, 865 (2006). 8. See Kenneth A. Bamberger, Provisional Precedent: Protecting Flexibility in Administrative Policymaking, 77 N.Y.U. L. REV. 1272, 1290–91 (2002); Sidney Shapiro & Elizabeth Fisher, Chevron and the Legitimacy of “Expert” Public Administration, 22 WM. &

MARY BILL RTS. REV. 465, 471–72 (2013). 9. See Jack M. Beermann, End the Failed Chevron Experiment Now: How Chevron Has Failed and Why It Can and Should Be Overruled, 42 CONN. L. REV. 779, 796 n.64 (2010) (reviewing literature); Lisa Schultz Bressman, Chevron’s Mistake, 58 DUKE L.J. 549, 554, 562 (2009); Antonin Scalia, Judicial Deference to Administrative Interpretations of Law, 1989 DUKE L.J. 511, 517. 10. See Beermann, supra note 9, at 795–96 (reviewing literature); Cynthia Farina, Statutory Interpretation and the Balance of Power in the Administrative State, 89 COLUM. L. REV. 452, 466–67 (1989); Scalia, supra note 9, at 515. 11. See Sanford N. Caust-Ellenbogen, Blank Checks: Restoring the Balance of Powers in the Post-Chevron Era, 32 B.C. L. REV. 757, 813–14 (1991); Meazell, supra note 3, at 1775–76 & n.57 (reviewing literature); Rajiv Mohan, Chevron and the President’s Role in the Legislative Process, 64 ADMIN. L. REV. 793, 795 (2012). But see Cass R. Sunstein, Beyond Marbury: The Executive’s Power to Say What the Law Is, 115 YALE L.J. 2580, 2582–83 (2006). 12. See John F. Duffy, Administrative Common Law in Judicial Review, 77 TEX. L. REV. 113, 189–211 (1998); Beermann, supra note 9, at 788–90 (citing Duffy and others). 13. See Shapiro & Fisher, supra note 8, at 471–72, 495. 14. See Brian G. Slocum, The Importance of Being Ambiguous: Substantive Canons, Stare Decisis, and the Central Role of Ambiguity Determinations in the Administrative State, 69 MD. L. REV. 791, 794 (2010); see also Scalia, supra note 9, at 520–21. 15. See Shapiro & Fisher, supra note 8, at 478.

2014] DEFERENCE IN STATE ADMINISTRATIVE LAW 557

law, has not been embraced with enthusiasm or consistency in state administrative law. This is true notwithstanding that state systems of administrative law in so many other key areas track their federal counterpart.16 I survey the many reasons why Chevron falls short of providing a useful framework for state judicial review. The executives, legislatures, courts, statutes, and bureaucracies of the states are different from those of the federal government in ways that do not suit Chevron. Taken individually, these differences are not overwhelming. They do not suggest that state administrative law in general should not bear a close family resemblance to its federal cousin. But they are significant enough that the rule of deference—the Chevron rule—is a poor candidate for policy diffusion into other jurisdictions, including the American states.

Ironically, the conclusion that Chevron is not the obvious model for deference doctrine in the states points not to a limitation of Chevron but to its great virtue. Chevron simultaneously addresses several different kinds of administrative-law problems, each with theoretical plausibility, pragmatism, and deeply contextualized analysis.17 The same multidimensionality that makes Chevron deeply potent as federal administrative law doctrine, even to its detractors, makes it less likely to fit the (only) somewhat different context of the states. Scholars like Evan Criddle, I suggest, are right to locate the ―genius‖ of Chevron in its ―pluralistic and conciliatory‖ character.18 The Chevron rule is bespoke doctrine, unsuited for drag-and-drop application even into other fairly similar political systems precisely because it is exquisitely attuned to its own context.

At the same time, Chevron‘s multivalence and contextualism are virtues that deserve and demand emulation by the states. States ought to mimic Chevron not by adopting its holding, but by thinking about what Chevron thinks about: how their legislatures operate, how their executives govern, how much scope their agencies have, what limits agency power, and what their agencies are capable of. State judges should seek, as Chevron sought, to find an approach to deference that is an equilibrium characterized by practical wisdom and that can coherently fit with the answers to all of these questions.

I. CHEVRON IN THE STATES

I do not undertake here to provide an updated survey of the reception of Chevron deference by state courts. Several high-quality and recent surveys

16. See Arthur Earl Bonfield, The Federal APA and State Administrative Law, 72 VA. L. REV. 297, 303 (1986). Bonfield notes that states did not copy federal administrative law; rather, the state and federal systems share common foundations. See id. at 298–99. 17. Evan Criddle, citing Rawls, calls this a ―pragmatic consensus.‖ Evan J. Criddle, Chevron’s Consensus, 88 B.U. L. REV. 1271, 1274 n.7, 1296 (2008). See also Sunstein, supra note 11, at 2596 (―The foundations of Chevron . . . are intensely pragmatic . . . .‖). 18. Criddle, supra note 17, at 1273.


are available.19 These surveys encounter the problem, surely unavoidable, that state courts‘ recital of deference rules does not seem always to reflect the actual nature of the deference that they provide agencies. (This phenomenon is known at the federal level as well.20) But the overall picture that these surveys provide is that of a ―mixed reception‖ for the Chevron rule in the state courts.

Some states are at extreme ends of the spectrum. Delaware‘s Public Water Supply v. DiPasquale21 is often cited as the most explicit repudiation of the Chevron framework in the states. The Delaware Supreme Court rejects Chevron on the grounds of precedent and because ―[s]tatutory interpretation is ultimately the responsibility of the courts.‖22 In Delaware, a ―reviewing court may accord due weight, but not defer, to an agency interpretation of a statute administered by it.‖23 The other state that rejects Chevron explicitly is Michigan, which repudiates earlier cases that had ―approvingly cited Chevron in the past‖24 for reasons similar to Delaware‘s:

The vagaries of Chevron jurisprudence do not provide a clear road map for courts in this state to apply when reviewing administrative decisions. Moreover, the unyielding deference to agency statutory construction required by Chevron conflicts with . . . the separation of powers . . . by compelling delegation of the judiciary‘s constitutional authority to construe statutes to another branch of government.25

On the other end of the spectrum are cases that adopt the Chevron rule. Professor Bernard Bell counts eleven states, plus the District of Columbia, that have adopted it explicitly.26 The Supreme Judicial Court of Maine, for example, states that in reviewing agency interpretation of ―a statute that an administrative agency administers and that is within its area of expertise,‖ it

19. See, e.g., WILLIAM N. ESKRIDGE, JR., PHILIP P. FRICKEY & ELIZABETH GARRETT, CASES AND MATERIALS ON LEGISLATION 1258–61 (4th ed. 2007) (―Note on Deference to Agencies in the State Courts‖); William R. Andersen, Chevron in the States: An Assessment and a Proposal, 58 ADMIN. L. REV. 1017, 1020–30 (2006); Bernard W. Bell, The Model APA and the Scope of Judicial Review: Importing Chevron into State Administrative Law, 20 WIDENER L.J. 801, 818–22 (2010); Ann Graham, Chevron Lite: How Much Deference Should Courts Give to State Agency Interpretation, 68 LA. L. REV. 1105, 1109–19 (2008); Michael Pappas, No Two-Stepping in the Laboratories: State Deference Standards and Their Implications for Improving the Chevron Doctrine, 39 MCGEORGE L. REV. 977, 980–1007, 1010–24 (2008). 20. See William N. Eskridge, Jr. & Lauren E. Baer, The Continuum of Deference: Supreme Court Treatment of Agency Statutory Interpretations from Chevron to Hamdan, 96 GEO. L.J. 1083, 1121 (2008). 21. 735 A.2d 378 (Del. 1999). 22. Id. at 382. The opinion also notes that ―it would be anomalous for this Court to accord a higher level of deference to the legal rulings of an administrative agency than that applied to trial courts subject to our appellate jurisdiction,‖ which are reviewed de novo. Id. at 381. The remaining analysis in the case focuses on Delaware precedents that address the question of deference. 23. Id. at 382. 24. In re Complaint of Rovas Against SBC Michigan, 754 N.W.2d 259, 271 n.63 (Mich. 2008). 25. Id. at 271–72. 26. See Bell, supra note 19, at 818 & n.108.


―determine[s] first whether the statute is ambiguous.‖27 If not, the court does ―not defer,‖ but ―interpret[s] the statute according to its plain language.‖28 But ―[i]f the statute is ambiguous, we defer to the agency‘s interpretation, and we affirm the agency‘s interpretation unless it is unreasonable.‖29 ―This,‖ says the Maine Supreme Court, ―is the same two-step analysis developed by the United States Supreme Court in Chevron.‖30 One might argue that Maine‘s limitation of the application of this doctrine to statutes within the agency‘s ―area of expertise‖31 gives the doctrine more limited application than Chevron enjoys in federal adjudication; but the doctrines are at least very close. Other states also appear to have adopted a doctrine that is similarly close to Chevron, but without citing the federal case explicitly.32

Most of the states, however, fall between the extremes of endorsing Chevron and repudiating it. Each survey of state doctrine notes a substantial number of states that ―defer‖ to agency determinations in the style of Skidmore v. Swift & Co.,33 i.e., to the extent that they are persuasive. Other states titrate the level of deference to match specific features of the agency at issue and the question at hand. This is similar to the ―continuum of deference‖34 that was given substantial momentum by dicta in United States v. Mead Corp.35 at the federal level.36 Broadly defined, this is the ―dominant view‖ among the states.37 Michael Asimow and his colleagues call those states that give agencies substantially less than full Chevron deference ―weak deference‖ states.38

The second most common approach is roughly to track Chevron, but to apply it in ways that give courts somewhat more, and agencies somewhat less, power. These states‘ doctrines strongly resemble Chevron but their courts neither cite it nor recapitulate its formulations.39 In particular, many

27. Cobb v. Bd. of Counseling Prof‘ls Licensure, 896 A.2d 271, 275 (Me. 2006). 28. Id. 29. Id. 30. Id. 31. Id. 32. See, e.g., Andersen, supra note 19, at 1025–26 (citing Rhode Island, Hawaii, and Mississippi as states that offer what ―seem to be boilerplate Chevron statements of the principle‖). 33. 323 U.S. 134 (1944). 34. Eskridge & Baer, supra note 20, at 1090. 35. 533 U.S. 218 (2001). 36. In his contribution to the symposium, Kent Barnett documents how the Congress has also begun to tailor the deference standard that applies in particular contexts—raising the question whether courts or legislatures are better situated to set the deference standard in particular situations. Kent Barnett, Improving Agencies’ Preemption Expertise with Chevmore Codification, 83 FORDHAM L. REV. 587 (2014) (detailing how in the Dodd-Frank Wall Street Reform and Consumer Protection Act, 12 U.S.C. § 25b(5)(A) (2012), the Congress directed courts to review under the Skidmore standard any decision to preempt state law made by the Office of the Comptroller of the Currency). 37. MICHAEL ASIMOW, ARTHUR EARL BONFIELD & RONALD M. LEVIN, STATE AND

FEDERAL ADMINISTRATIVE LAW § 9.2, at 563 (2d ed. 1998); see also id. at 557–58. 38. Id. at 557–58. 39. See, e.g., David E. Shipley, The Chevron Two-Step in Georgia’s Administrative Law, 46 GA. L. REV. 871, 923 (2012) (stating that although Georgia applies a deference regime


states do not adopt the two-step aspect of Chevron, declining to articulate two discrete inquiries, one de novo and one deferential. The literature at the federal level demonstrates that this may or may not be a distinction without a difference: Controversy swirls there regarding whether Chevron itself is best understood as having only one step,40 or, as some commentators on the state cases have also argued (with varying degrees of convincingness), whether the two-step structure carries doctrinal and practical significance.41

It is also fairly common for states to articulate a doctrine similar to Chevron‘s but then accord deference more stingily than Chevron, honestly applied, seems to require. Again, this phenomenon has a cognate in the federal system.

II. CONSTITUTIONAL INSTITUTIONS

Why has Chevron, so dominant in federal law, found relatively little purchase in the states? To the extent that Chevron is a case about separation of powers, delegation, and political theory—these categories are not easily disentangled, and for my purposes disentanglement is not necessary—one critical reason might be that state constitutions and their conceptions of separation of powers are unlike their federal counterparts. This part reviews the key differences between state constitutional institutions and the federal constitutional structure in which Chevron finds much of its justification.42

A. State Judges

Chevron is fundamentally a judicially articulated restriction upon judicial power.43 In this sense it rests upon federal separation of power doctrine, which understands judges as apart from, and under an obligation to respect, the work of ―the political branches.‖44 Chevron contends that this work

similar to Chevron‘s, ―there is no statement from either the Georgia Supreme Court or the Georgia Court of Appeals as clear as that key paragraph from Chevron announcing the United States Supreme Court‘s two-step approach‖). 40. Compare Matthew C. Stephenson & Adrian Vermeule, Chevron Has Only One Step, 95 VA. L. REV. 597 (2009), with Kenneth A. Bamberger & Peter Strauss, Chevron’s Two Steps, 95 VA. L. REV. 611 (2009). See also Richard M. Re, Should Chevron Have Two Steps?, 89 IND. L. REV. 605, 606–08 (2014) (reviewing literature). 41. See, e.g., Pappas, supra note 19, at 980, 986, 995 (―[T]here is an important difference between Chevron, as currently practiced, and [some state courts‘] equivalent one-step reasonableness test‖); see also Scott A. Keller, Texas Versus Chevron: Texas Administrative Law on Agency Deference After Railroad Commission v. Texas Citizens, 74 TEX. B.J. 984, 986 (2011). 42. Jim Rossi calls these kinds of differences ―imperfections‖ in state constitutional law, by which he means departures from the federal constitutional paradigm rather than errors in state constitutional design. See Jim Rossi, Overcoming Parochialism: State Administrative Procedure and Institutional Design, 53 ADMIN. L. REV. 551, 557–59 (2001) [hereinafter Rossi, Overcoming Parochialism]. 43. See Thomas W. Merrill, Textualism and the Future of the Chevron Doctrine, 72 WASH. U. L.Q. 351, 353 (1994). 44. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 866 (1984) (quoting Tenn. Valley Auth. v. Hill, 437 U.S. 153, 195 (1978)).


includes the interpretation of ambiguity. Such interpretation often implicates questions of ―policy,‖ questions ―more properly addressed to legislators or administrators, not to judges.‖45 Federal judges ―have no constituency‖; for this reason they ―have a duty to respect legitimate policy choices made by those who do.‖46

Chevron thus justifies its rule not only upon grounds of legislative and executive accountability but also of judicial nonaccountability. It matters that federal judges lack a political constituency. The combination of the popular accountability of the elected branches and the lack of parallel accountability in the judiciary justifies Chevron‘s conclusion that an ambiguous administrative statute is an implicit delegation of interpretative authority to the agency rather than a problem that ought to be solved in the courts.47

Aaron-Andrew Bruhl and Ethan Leib recently offered a thorough and cogent argument for the proposition that state judges differ from federal judges in ways that potentially undermine the applicability of Chevron to the state context.48 The most obvious difference, they suggest, is that state judges are not appointed for life.49 State arrangements are all over the map: some judges in some states are elected, some appointed but subject to reappointment, some appointed but subject to retention elections, some appointed for fixed terms, and some subject to mandatory retirement.50 Judges subject to such arrangements in an important sense do have political constituencies. If judges face an electorate, or if their tenure is subject to retention decisions by elected officials, they are in an important sense ―political creatures.‖51 As Bruhl and Leib point out, this means that the courts on which they sit are in an important sense ―political branches.‖52

Bruhl and Leib‘s argument goes beyond the simple observation that elected judges are politicians because they are elected. Elections do more than make state judges democratically accountable; they also make them politically attuned and politically connected. By virtue of elections and set terms, they are closer to the public than federal judges; by virtue of the

45. Id. at 864. 46. Id. at 866. 47. See Sunstein, supra note 11, at 2586. 48. Aaron-Andrew P. Bruhl & Ethan J. Leib, Elected Judges and Statutory Interpretation, 79 U. CHI. L. REV. 1215, 1249–54, 1277–82 (2012). 49. Id. at 1250. 50. See LARRY BERKSON, SCOTT BELLER & MICHELE GRIMALDI, JUDICIAL SELECTION IN

THE UNITED STATES: A COMPENDIUM OF PROVISIONS 6 (1980) (―Today there is an almost endless combination of schemes used to select judges. Almost no two states are alike . . . .‖); Joseph A. Colquitta, Rethinking Judicial Nominating Commissions: Independence, Accountability, and Public Support, 34 FORDHAM URB. L.J. 73, 73, 77 (2007); F. Andrew Hanssen, The Effect of Judicial Institutions on Uncertainty and the Rate of Litigation: The Election Versus Appointment of State Judges, 28 J.L. STUD. 205, 211 (1999); Lawrence Schlam, State Constitutional Amending, Independent Interpretation, and Political Culture: A Case Study in Constitutional Stagnation, 43 DEPAUL L. REV. 269, 293 & n.60 (1994). 51. Bruhl & Leib, supra note 48, at 1253. 52. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 838 (1984); see Bruhl & Leib, supra note 48, at 1249.


political nature of their office they are also closer to the legislature.53 State judges are thus potentially, in at least some important contexts, differently situated than federal judges with respect to understanding and interpreting statutes that grant agencies power.54 All of these factors weaken the case that state judges are like federal judges with respect to the need to cede ―policy‖ determination to other branches.

Professor Bernard Bell likewise notes that various aspects of state court jurisdiction make state judges more like policymakers than their federal counterparts.55 State courts have the general jurisdiction to develop the common law.56 They may render advisory opinions.57 They have generous standing and justiciability rules.58

What is to be made of these differences? One possibility is that elected judges, both more politically attuned and more politically vulnerable than their appointed counterparts, might be particularly activist with respect to judicial review. Where an appointed judge sticks to the law, an elected judge might, like any other politician, more readily seek to give expression to ―policy‖ views. At the extreme, the elected judge may find herself beholden to or even ―controlled‖ by special interests or political machines.59

F. Andrew Hanssen, by contrast, hypothesizes the converse, that appointed judges can oversee agencies with ―greater impunity‖ than elected ones.60 He suggests that ―appointed courts may be expected to be less influenced by the political/electoral forces that underlie the policy decisions of administrative agencies (at least to a degree), and, accordingly, to be more threatening to agency decisions.‖61

William Funk, and Bell after him, suggests the third possibility that elected and appointed judges are not situated differently with respect to judicial review.62 In Funk‘s words:

The fact that many state judges are subject to the electoral process also seems irrelevant to their suitability to review the rationality of state agency rulemaking. Or at least it is no more relevant to their suitability to

53. See Bruhl & Leib, supra note 48, at 1250–54. Abbe Gluck extends this observation to state judges in general, regardless whether they are elected. See Abbe R. Gluck, The States As Laboratories of Statutory Interpretation: Methodological Consensus and the New Modified Textualism, 119 YALE L.J. 1750, 1816 (2010). 54. But see William Funk, Rationality Review of State Administrative Rulemaking, 43 ADMIN. L. REV. 147, 161 n.95 (1991) (―In many states . . . judges are elected or subject to reelection, but no one has suggested that courts are intended to be politically responsive in these states.‖). 55. Bell, supra note 19, at 824–25. 56. Id. 57. Id. at 825. 58. Id. at 825–26. 59. BERKSON, BELLER & GRIMALDI, supra note 50, at 4. 60. F. Andrew Hanssen, Independent Courts and Administrative Agencies: An Empirical Analysis of the States, 16 J.L. ECON. & ORG. 534, 538 (2000). 61. Id. at 535. Hanssen tests this hypothesis by looking for correlations between judicial elections and agency staffing levels for agencies subject to litigation. See id. at 535–36. 62. See Bell, supra note 19, at 827–28; Funk, supra note 54, at 174–75.


engage in rationality review of rulemaking than it is to their suitability to engage in any other form of judicial activity.63

Finally, Bruhl and Leib, for their part, understand institutional variation in state judiciaries to imply ―that one has to embrace nuance when thinking about when deference is indicated in the state courts.‖64 They do not claim that Chevron is never appropriate in the state context. Nor do the authors mean by ―nuance‖ just that one might adopt different modes of deference depending upon the constitutional structure of a state judiciary, whether it is elected or appointed, term-limited or not. Bruhl and Leib mean nuance: that the regime of judicial deference should depend upon the kind of electoral accountability a particular court faces. Judicial elections, they suggest, create too many different kinds of situations for a rule like Chevron to have broad applicability.

From a positive perspective, scholars have been unable to demonstrate a clear relationship between variables related to the election and retention of state judges and their decisions in cases involving the executive branch.65 The ―somewhat mixed‖66 empirical results are consistent both with the possibility that judicial election and appointment do not matter and the possibility that nuance, which this sort of analysis cannot capture, rules the day.

But, of course, nuance and Chevron are uneasy bedfellows. Bruhl and Leib use the electability of judges as an argument for the ―continuum of deference,‖67 identified at the federal level with Mead.68 And although Mead does not repudiate Chevron, this continuum is at odds with the Chevron idea that structural principles dictate a one-size-fits-all deference regime. Justice Scalia, dissenting in Mead, is Chevron‘s great champion because he rejects a continuum of deference. To ―tailor deference to variety‖69 is to limit, not to endorse, the scope of Chevron.

B. State Executives

Chevron also rests heavily upon the relationship between the legislature and the executive, and the accountability of agencies to the latter. Judicial acquiescence to reasonable agency interpretation is appropriate, Chevron tells us, because unlike judges,

an agency to which Congress has delegated policy-making responsibilities may, within the limits of that delegation, properly rely upon the

63. Funk, supra note 54, at 174–75. 64. Bruhl & Leib, supra note 48, at 1279 & n.245. 65. See Gbemende Johnson, Judicial Deference and Executive Control over Agencies, 14 ST. POL. & POL‘Y Q. 142, 143–44 (2014). The measure that analysts generally use for deference in these studies is the rate at which the executive branch prevails in cases where it is a party. This is obviously dramatically different from, but has some difficult-to-specify relationship with, the question of the appropriate doctrinal framework for deference. 66. Id. at 144. 67. See Eskridge & Baer, supra note 20, at 1098. 68. United States v. Mead Corp., 533 U.S. 218 (2001). 69. See id. at 236 (―Justice Scalia‘s first priority over the years has been to limit and simplify. The Court‘s choice has been to tailor deference to variety.‖).


incumbent administration‘s views of wise policy to inform its judgments. While agencies are not directly accountable to the people, the Chief Executive is, and it is entirely appropriate for this political branch of the Government to make such policy choices—resolving the competing interests which Congress itself either inadvertently did not resolve, or intentionally left to be resolved by the agency charged with the administration of the statute . . . .70

This statement justifiably has been the source of considerable controversy. It troublingly and casually conflates the President, the ―administration,‖ and the agencies.71 Even more problematically, it seems to assume routine, hierarchical control by the first two over the third. But, holding aside whether Chevron properly describes the presidency and its relationship to the agencies, it seems incontrovertible that federal agencies are embedded in some fashion in the executive branch, in ways that create political accountability that is more direct than that of federal courts. Federal agency heads are clearly subject to the political supervision of the President or his ―administration‖ in some fashion: they face presidential appointment, the requirements of the Office of Information and Regulatory Affairs (OIRA), and, for line agencies, the possibility of presidential removal.

All three of these institutional structures—appointment, review, and removal—exist in the states.72 But the structural and institutional context of the states is not Washington‘s. The assumption of hierarchical control in particular is even more implausible with respect to the states than it is with respect to the federal government. ―Despite numerous state reorganizations,‖ Jeffrey Brudney and F. Ted Hebert report, state executive branches are ―not hierarchical organization[s] yielding to classical management principles with the governor securely in control. In fact, great

70. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 865–66 (1984). 71. Substantial controversy continues to swirl over whether and to what extent agencies are ―directly accountable‖ to the President. ―Unitarian‖ theorists who argue that agencies are so accountable, see Christopher C. DeMuth & Douglas H. Ginsburg, White House Review of Agency Rulemaking, 99 HARV. L. REV. 1075, 1083 (1986); Paul R. Verkuil, Jawboning Administrative Agencies: Ex Parte Contacts by the White House, 80 COLUM. L. REV. 943, 957 (1980), are likely to be less troubled by Chevron‘s reliance upon the popular accountability of the person of the Chief Executive than those who champion the position that agencies are not subject in all things to direct line supervision by the President. See, e.g., Peter M. Shane, Chevron Deference, the Rule of Law, and Presidential Influence in the Administrative State, 83 FORDHAM L. REV. 679 (2014); Kevin M. Stack, The President’s Statutory Powers To Administer the Laws, 106 COLUM. L. REV. 263, 299 (2006); Peter L. Strauss, Overseer, or “The Decider”? The President in Administrative Law, 75 GEO. WASH. L. REV. 696, 759–60 (2007); see also Elena Kagan, Presidential Administration, 114 HARV. L. REV. 2245, 2289 (2001) (characterizing this view as one that allows ―the President to issue only an advisory opinion‖). 72. See Neal D. Woods & Michael Baranowski, Governors and the Bureaucracy: Executive Resources As Sources of Administrative Influence, 30 INT‘L J. PUB. ADMIN. 1219, 1221 (2007) (arguing that appointment is an ―important‖ and ―fundamental‖ source of gubernatorial influence over agency action).


variation distinguishes the degree to which agencies are subordinate—formally or informally—to the governor.‖73

In the executive sphere, the primary structural difference is that most states have plural, rather than unitary, executives. Christopher Berry and Jacob Gersen have documented the extraordinary extent of plural executive arrangements across the states. These institutions are startling both because of the number of elected statewide officials and the apparent trend in favor of what Berry and Gersen call executive ―unbundling.‖74 In 2002, they report, governors and lieutenant governors were elected independently of one another in forty-five states.75 Treasurers and secretaries of state were elected independently in thirty-eight.76 Berry and Gersen identified more than five states that, variously, use direct election for the chief executives of state agencies that govern auditing, education, agriculture, insurance, utilities, community affairs, and finance.77 In the states, then, agencies in some spheres are directly accountable to the people, and the governor, in those spheres, is not.78

This need not undermine Chevron-style deference. Funk argues, for example, that the popular election of agency heads should not be relevant to the doctrine of judicial review. Such agencies are still engaged in rulemaking and administrative adjudication, not in legislation, much as popularly elected judges are engaged in adjudication notwithstanding election. Instead, agency action should be reviewed under the assumption that administrative rather than legislative norms govern its production.79

Indeed, one might make an even stronger argument. As Berry and Gersen have argued, if particular agency heads are subject to direct public election, there is every reason to think that policy in the domain belonging to that agency will be determined in closer accordance with popular will than it would be were that agency subordinated to a ―general purpose‖ chief executive.80 An entire layer of principal-agent problems is avoided.81 On this argument, perhaps Chevron applies a fortiori to agencies directly elected and directly accountable to the people: when the legislature delegates authority to such an agency, and legislative authority proves to have explicit or implicit gaps, the agency‘s direct accountability provides an especially strong argument for presuming that those gaps constitute a

73. Jeffrey L. Brudney & F. Ted Hebert, State Agencies and Their Environments: Examining the Influence of Important External Actors, 49 J. POL. 186, 188, 192–94 (1987) (citations omitted). 74. See generally Christopher R. Berry & Jacob E. Gersen, The Unbundled Executive, 75 U. CHI. L. REV. 1385 (2008). For an overview of nonunitary executives in the states, see Rossi, Overcoming Parochialism, supra note 42, at 557–59. 75. Berry & Gersen, supra note 74, at 1433 tbl.4. 76. Id. 77. Id. 78. As an historical matter, the trend in the states is clearly in the direction of concentrating ultimate executive power in fewer officials. See JON C. TEAFORD, THE RISE OF

THE STATES: EVOLUTION OF AMERICAN STATE GOVERNMENT 204–05 (2002). 79. Funk, supra note 54, at 178–79. 80. Berry & Gersen, supra note 74, at 1393–94. 81. Id. at 1394.


delegation of law-determining power to that agency, within the bounds of reasonableness.

But it is more complicated than that. For one thing, election is not the same thing as accountability. Like special-purpose local governments and single-topic congressional committees, a state agency with an elected head likely is less politically salient than a general-purpose executive—a governor or lieutenant governor. Elections for special-purpose state officials are substantially less vivid and relevant to their electorates than national politics (although there are dramatic exceptions82). When Chevron refers to the political accountability of the ―Chief Executive,‖83 it contemplates an executive, the President, whose name everyone knows. Likewise, the state-level electorate is familiar with its governor, and blames him for things it does not like. This is true even if credit and blame properly belong to the elected state insurance or agriculture commissioner, persons unknown to almost everyone.84 For the same reasons, the special-purpose elected state official seems to be a prime candidate for capture by special interests.85

Second, plural executive arrangements give rise to various kinds of intraexecutive checks. Of particular interest are state attorneys general, who are elected, rather than appointed by governors, in forty-three states.86 Scott M. Matheson, Jr. notes that in some states ―the attorney general has the exclusive constitutional authority to employ and supervise all lawyers involved in legal matters for executive agencies of state government, including the lawyers for other elected executive branch officials.‖87 This might allow an attorney general‘s interpretation of the law to trump that of both the governor and the agency head. At a minimum ―the problem of deadlock and frustration of executive branch mission is a real possibility,‖ especially, Matheson adds, when there are partisan differences.88

In addition, in many (but not all) states, attorneys general may and do bring suit against other elected executive officials, at least in cases of potential unconstitutionally.89 William Marshall outlines an argument that attorneys general are not supposed to substitute their policy judgment for that of agencies but are supposed to exercise independent legal judgment.90 This argument is especially strong, Marshall suggests, when attorneys

82. For a recent example, see David Montgomery, Slowed Elsewhere, Tea Party Still Wields Considerable Sway in Texas Races, N.Y. TIMES, May 26, 2014, at A9. 83. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 865–66 (1984). 84. See LARRY J. SABATO, GOOD-BYE TO GOOD-TIME CHARLIE 63–65 (2d ed. 1983). 85. Cf. Steven G. Calabresi & Nicholas Terrell, The Fatally Flawed Theory of the Unbundled Executive, 93 MINN. L. REV. 1696, 1700 (2009). 86. Berry & Gersen, supra note 74, at 1433 tbl.4. 87. See Scott M. Matheson, Jr., Constitutional Status and Role of the State Attorney General, 6 U. FLA. J.L. & PUB. POL‘Y 1, 18, 21 (1993). 88. See id. at 18. 89. William P. Marshall, Break Up the Presidency? Governors, State Attorneys General, and Lessons from the Divided Executive, 115 YALE L.J. 2446, 2458–60 (2006). These suits can become complicated. See Matheson, supra note 87, at 25. 90. See Marshall, supra note 89, at 2464.


general sue another entity within the ―executive branch for exceeding its authority.‖91 Such suits, of course, are quintessential Chevron analogs.

In accordance with Marshall‘s ―structural‖ argument, and on one reading of Chevron, attorneys general are exercising their own brand of Chevron deference. Chevron tells courts to defer to agencies making ―policy choices [] resolving . . . competing interests,‖92 and the structural argument tells attorneys general to do the same thing. But of course Chevron understands the ―policy choices‖ to which it refers to be instantiated as interpretations of the law; and, as Marshall argues, the law is the proper subject for the attorney general‘s independent judgment. In that respect, the state attorney general is a Chevron-confounding figure. He or she constitutes a second source within the politically accountable executive that independently determines what the law is in the case of a legislative gap.

The potential such arrangements create for intraexecutive conflicts can push judicial deference in a variety of conflicting directions. On the one hand, the plural executive is a mechanism for checking and balancing unitary executive power. Judicial review, which is also supposed to check that power, might therefore feel itself able to be quite deferential, because its check is less vital. If the attorney general and an agency agree, perhaps state courts should provide full Chevron-style deference, or be even more deferential. But if they disagree, Chevron provides little guidance: Should one defer to the expert agency, as Chevron urges, or the more accountable state attorney general, as Chevron also urges? In cases of disagreement, moreover, and perhaps some cases of agreement as well, multiple executives ―will have powerful incentives to blur accountability by blurring the distinctions between their own power and the power of their coexecutive colleagues.‖93 Indeed, plural executives make it difficult to assign responsibility regardless whether such blurring is desired.94

And what of an expert agency whose chief is directly elected, rather than subordinate to the governor? A state-elected education or insurance commissioner, in conflict with an attorney general or governor or both, is both accountable (because she is elected) and expert (because she is an agency with a particular policy portfolio). But the cases for both her accountability and her expertise could potentially be quite weak. She might be elected in a low-salience or very low-salience election, much less accountable than the attorney general or governor with whom she is at odds. And she might also be, in part because of the fact of her election, inexpert as well; certainly she could well be less expert than the technocratic bureaucracies that Chevron postulates are working in the federal context.

As with the state judiciary, therefore, we come to a position of doubt with respect to the Chevron framework as applied to the state executive. There will surely be states, and particular categories of institutional and factual circumstances within those states, where Chevron‘s arguments about

91. Id. at 2466. 92. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 865 (1984). 93. Calabresi & Terrell, supra note 85, at 1712. 94. See Matheson, supra note 87, at 18.


executive accountability seem as apropos as they do in the federal system. But there will surely be many others where they seem much more inapposite than they do in the federal system. The appeal of Chevron as a general principle, then, fades in the face of the widespread phenomenon of the plural state executive and intraexecutive conflict.

C. State Legislatures

Chevron‘s principal structural justification for its deference rule is that gaps in agency-empowering legislation constitute a ―legislative delegation to an agency on a particular question.‖95 The cognate state institution, the state legislature, is therefore critical with respect to the applicability of Chevron deference. In particular, Chevron‘s innovation was to discern in legislative gaps ―implicit‖ delegations that paralleled explicit ones. Both are ―delegations.‖ It is therefore important to consider whether state legislatures, regardless whether the gaps they leave are implicit or explicit, plausibly can be thought to be ―delegating‖ power to agencies in the same sense as does the Congress of the United States.

There is some reason to think so. Of the three branches of government, state legislatures are structurally most like their federal counterparts. All but one state legislature is bicameral.96 All state legislatures, like the Congress, enact legislation that undertakes to guide agency action.97 Institutions like committees, markup, and reconciliation, though they often differ in their details and in their effectiveness, are common across states and the federal system.98

But there are differences that are arguably relevant. Perhaps the most important is the peculiar insistence in the federal courts upon the principle that the Congress may not delegate any legislative power to agencies (―nondelegation‖).99 But, in federal law, authorizations to adjudicate or to make rules that guide agency discretion are legitimate delegations of executive power, rather than illegitimate ones of legislative power, so long as the agency‘s discretion is cabined by an ―intelligible principle.‖100 The requirement of intelligibility in the United States Supreme Court is quite undemanding.101

As Jim Rossi has exhaustively documented, this is not the rule of delegation in the states.102 Rossi finds states that locate themselves on both

95. Chevron, 467 U.S. at 844. 96. THE LEGISLATIVE BRANCH OF STATE GOVERNMENT: PEOPLE, PROCESS, AND

POLITICS 6 (Thomas H. Little & David B. Ogle eds. 2006). 97. Id. at 2, 11. 98. Id. at 7; Richard A. Briffault, Beyond Congress: The Study of State and Local Legislatures, 7 N.Y.U. J. LEGIS. & PUB. POL‘Y 23, 25 (2003). 99. Whitman v. Am. Trucking Ass‘ns, Inc., 531 U.S. 457, 472 (2001). 100. Id. 101. See id. at 474. 102. Jim Rossi, Institutional Design and the Lingering Legacy of Antifederalist Separation of Powers Ideals in the States, 52 VAND. L. REV. 1167 (1999) [hereinafter Rossi, Institutional Design]; see also Gary J. Greco, Standards or Safeguards: A Survey of the Delegation Doctrine in the States, 8 ADMIN. L.J. AM. U. 567, 579–80 (1994).


sides of the federal doctrine. A handful of state courts will bless even agency delegations that contain no substantive restrictions upon agency discretion, so long as procedural safeguards are in place.103 A much larger group is less tolerant of legislative delegation than federal law. Some of these put teeth into the concept of intelligibility not present in the federal doctrine.104 Others understand the power to set the broad outlines of important policies to be legislative, and nondelegable, even if the legislature offers the agency some intelligible guide to the exercise of its discretion.105

If we understand Chevron‘s idea of implicit delegation in the context of federal nondelegation doctrine,106 then the implications of state nondelegation doctrine for Chevron review could be substantial. The ―implicit‖ delegation by the Congress to the agency that Chevron sees in the passage of an ambiguous act that does not ―directly‖ speak ―to the precise question at issue‖107 sets up statutory ambiguity. This ambiguity, which limns the boundaries of reasonableness in the Step-Two inquiry, validates the delegation.108 If an ambiguous grant of authority is written to make more than one but not every possible agency interpretation of its authority reasonable, then its principle is intelligible.

But if a particular state does not require the presence of an intelligible principle for delegation, then the presence of an ambiguity less plausibly indicates that the legislature intended, by the fact of ambiguity, such a delegation. Similarly, if a particular state requires more than the presence of an intelligible principle for licit delegation, the mere fact of ambiguity again seems a less convincing basis upon which to imply a delegation. In both kinds of states—together, easily most states—the Chevron presumption that ambiguity constitutes delegation seems weak. Ambiguity in these states seems rather to be just ambiguity: a statement of the law that will eventually require someone to declare what the law is. Without the delegation argument, it becomes more plausible that this someone should be a court reviewing de novo.

This argument is further buttressed by other structural differences between the Congress and many state legislatures. State legislatures have general jurisdiction; unlike the Congress, they may legislate about virtually any topic not preempted by federal law.109 The scope of potential delegation is therefore much greater in the state than in the federal

103. Rossi, Institutional Design, supra note 102, at 1191–93. 104. Id. at 1193–94. Rossi also notes that several state courts whose formulation of their nondelegation doctrines parallels that of the U.S. Supreme Court in practice appear to have somewhat more demanding standards than the federal courts for what makes a principle sufficiently intelligible to qualify as not delegating legislative power. Id. at 1200. 105. Id. at 1195–97. 106. See Patrick M. Garry, Accommodating the Administrative State: The Interrelationship Between the Chevron and Nondelegation Doctrines, 38 ARIZ. ST. L.J. 921, 923 (2006) (―The evolution of the [federal] nondelegation doctrine essentially necessitates the Chevron doctrine.‖). 107. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842 (1984). 108. See Shane, supra note 71. 109. See Rossi, Overcoming Parochialism, supra note 42, at 566.


legislature. General jurisdiction is likely one reason for more rigorous nondelegation doctrine in the states. Similar considerations could lead to the conclusion that one ought to be hesitant to infer delegation by state legislatures from the mere fact of statutory ambiguity.

In other work, Jim Rossi has also noted other, less substantial but still salient differences between the Congress and the legislatures. Legislatures, often pursuant to constitutional provisions, meet infrequently.110 In some states serving as a legislator is a part-time job and legislative staffs are less developed.111 Infrequent sessions, high turnover, and limited staff, argues Rossi, mean less internal expertise, a greater scope for lobbyists, and, overall ―decreased legislative accountability.‖112 State legislatures are also less in a position to oversee agency activity through staff work and oversight hearings, which are ubiquitous management devices at the federal level.113 All of these factors suggest that there is less reason at the state than at the federal level to see in an ambiguous statute a legislative intent to delegate power to agencies.

III. STATUTES GOVERNING REVIEW OF AGENCY ACTION

Chevron famously does not cite the Administrative Procedure Act.114 Nor by any means do the provisions of the APA imply or require Chevron (or any other) kind of deference to agencies‘ statutory interpretations. Indeed, some scholars argue that Chevron contravenes the APA, which they understand to require courts to review agencies‘ legal interpretations de novo.115

An opposing view understands Chevron to be careful to align itself with statutory principles of judicial review. Chevron begins with the principle that courts must give agency constructions pursuant to explicit delegation of executive power ―controlling weight unless they are arbitrary, capricious, or manifestly contrary to the statute.‖116 It then extends this principle to implicit delegations. In doing so, Chevron cites to Court precedent—not to the APA—but its language parallels117 the APA requirement that reviewing courts ―hold unlawful and set aside agency action‖ that is found to be ―arbitrary, capricious, . . . or otherwise not in accordance with law.‖118 The APA does not mandate deference to agencies‘ legal interpretations; it is consistent with the statute for courts to determine de novo what is or is not ―in accordance with law.‖ But Chevron‘s is nonetheless an APA-

110. See Briffault, supra note 98, at 27. 111. See id. at 25–26. 112. See Rossi, Overcoming Parochialism, supra note 42, at 560. 113. Id. at 567. 114. See Sunstein, supra note 11, at 2586. 115. See, e.g., Beermann, supra note 9, at 788–94; Duffy, supra note 12, at 193–98; Elizabeth Garrett, Legislating Chevron, 101 MICH. L. REV. 2637 & n.17 (2003); Sunstein, supra note 11, at 2590. 116. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 844 (1984). 117. See Steve R. Johnson, Preserving Fairness in Tax Administration in the Mayo Era, 32 VA. TAX REV. 269, 283 n.74 (2012). 118. 5 U.S.C. § 706(2) (2012).


compatible framework for evaluating such interpretations, one that is consistent with, and fits cleanly into, the APA‘s conceptual framework for judicial review.

Regardless which of these views one accepts, Chevron fits less neatly into the framework of state-level Administrative Procedure Acts (SLAPAs) than into that of the federal APA. This is true in two major respects.119 First, Chevron tracks SLAPA language regarding judicial review, which is itself variable, less cleanly than it does federal language. Second, most states establish by statute, along with executive order, mechanisms for the review of agency rules by legislatures and governors as well as by courts. The review of agency action by nonjudicial actors perforce carries implications for review conducted by judges.

A. Statutory Language Regarding Judicial Review

The 2010 Model State Administrative Procedure Act (MSAPA) directs courts to provide relief if ―agency action is arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.‖120 A separate, independent, and prior provision also directs courts to provide relief if ―the agency erroneously interpreted the law.‖121 These provisions together parallel the federal APA more closely than did the earlier, 1981 version of the Model State APA, which directed state courts to set aside agency action that was ―outside the range of discretion delegated to the agency by any provision of law‖ and, in brackets, also suggested that state courts set aside agency action ―otherwise unreasonable, arbitrary, or capricious.‖122 ―The brackets,‖ explained Professor William Funk in 1991, ―indicate the Uniform Law Commissioners‘ intent to provide an option between a broader and more limited judicial role in review of agency action.‖123

The current Model Act, in elevating the familiar arbitrary and capricious standard to a more central position, is perhaps somewhat less catholic than previous versions of the Act with respect to this question; by paralleling federal language it can fairly be said to have moved in the direction of Chevron-style deference. But this is at most a small change.124 The current Model Act, like its predecessor, remains determinedly and purposefully

119. A third issue is that some states have general statutes governing judicial statutory construction that have no federal analog and arguably conflict with Chevron. See, e.g., Manuel H. Hernandez, Running Out of Gas: Why Texas Must Distance Itself Completely From the Chevron Doctrine of Administrative Deference, 14 TEX. TECH. ADMIN. L.J. 225, 229–30 (2012) (citing TEX. GOV‘T CODE ANN. § 311.023 (West 2005)). 120. REVISED MODEL STATE ADMIN. PROCEDURE ACT § 508(a)(3)(C) (2010). 121. Id. § 508(a)(3)(A). 122. See MODEL STATE ADMIN. PROCEDURE ACT § 5-116(c)(8) (1981) (reproduced at ASIMOW ET AL., supra note 37, at 777–78). Note that the directive to set aside agency action ―outside the range of discretion delegated to the agency by any provision of law‖ eschews Chevron‘s distinction between agencies‘ interpretations of their own organic statutes and those of other statutes. 123. See Funk, supra note 54, at 154. 124. See John Gedid, Modernizing Agency Practice: The 2010 Model State Administrative Procedure Act, 20 WIDENER L.J. 697, 704 (2011) (―One area where there is little change in the 2010 MSAPA is judicial review.‖).


unwilling to endorse Chevron deference.125 The drafters comment that the ―arbitrary, capricious, . . . or otherwise not in accordance with law‖ provision ―is not intended to preclude courts from according deference to agency interpretations of law, where such deference is appropriate.‖126 However, they also cite approvingly William Araiza‘s view that the ―[j]udiciary, not [the] legislature, [is the] appropriate body to evolve specific standards for review, because of [the] great variety of agency action and contexts.‖127 In addition, the introduction to the 2010 Act, after noting that ―[m]ost states have a substantial body of judicial review case law,‖ states that its ―provisions are designed to be consistent with the existing laws of many states that take a variety of approaches to judicial review.‖128 Moreover, the separate provision requiring relief in the case of erroneous agency interpretation remains. By separating law-interpretation cases from other kinds of agency action, the Model Act leaves plenty of room for separate standards.

Michael Pappas‘s discussion of deference demonstrates both that state courts appear to be concerned that their regimes of deference be conceptually consistent with SLAPA requirements129 and that not all SLAPAs bear as much of a family resemblance to the federal Act as the MSAPA.130 For example, he argues that Public Water Supply Co. v. DiPasquale,131 the Delaware case that repudiates Chevron in favor of de novo review, is rooted in the Delaware SLAPA. De novo review is not compelled by that state act, which is silent with respect to standards of deference to agencies‘ statutory interpretation.132 But the Delaware Act combines that silence with a requirement that courts defer to agencies‘ factual findings,133 which suggests an expressio unius argument that deference should not be accorded to the former. Moreover, the Delaware Act requires agencies reviewing regulations to ―take due account of the experience and specialized competence of the agency and of the purposes of the basic law under which the agency acted.‖134 This can be read to suggest that (a) inquiry as to legal meaning must precede the consideration of agency action, and (b) Chevron‘s blanket deference to agencies by virtue of their institutional identity as agencies making decisions that carry the force of law is not what the framers of the Delaware Act had in mind.

Another way in which SLAPAs address the deference problem is by instructing agencies to read their authorizing statutes narrowly. Professor Bell cites the excellent examples of the Colorado and Florida SLAPAs.

125. Bell, supra note 19, at 821. 126. REVISED MODEL STATE ADMIN. PROCEDURE ACT § 508 cmt. (2010). 127. Id. (citing William D. Araiza, In Praise of a Skeletal APA, 56 ADMIN. L. REV. 979 (2004)). 128. Id. at 5 (prefatory note). 129. Pappas, supra note 19, at 978–79. 130. Id. at 984–87; see also Funk, supra note 54, at 154–55. 131. 735 A.2d 378 (Del. 1999). 132. Pappas, supra note 19, at 990. 133. Id. 134. DEL. CODE ANN. tit. 29, § 10141(e) (2003) (cited in Pappas, supra note 19, at 989).


The Colorado statute states that ―[n]o rule shall be issued except within the power delegated to the agency and as authorized by law,‖ and explicitly adds that a ―rule shall not be deemed to be within the statutory authority and jurisdiction of any agency merely because such rule is not contrary to the specific provisions of a statute.‖135 Florida‘s statute states that

[a]n agency may adopt only rules that implement or interpret the specific powers and duties granted by the enabling statute. No agency shall have authority to adopt a rule only because it is reasonably related to the purpose of the enabling legislation and is not arbitrary and capricious or is within the agency‘s class of powers and duties, nor shall an agency have the authority to implement statutory provisions setting forth general legislative intent or policy.136

These statutes and cognate provisions in other states do not reject Chevron. Rather, they are provisions that guide statutory interpretation by agencies. Just as Chevron itself instructs the federal courts to address Step One problems using ―the traditional tools of statutory construction,‖137 agencies faithful to the textualist instructions of the Colorado or Florida legislatures could still encounter ambiguities permitting multiple interpretations, and courts could still then defer to their reasonable choices among those interpretations. But these sorts of statutes surely do restrict the potential reach of any state-level doctrine cognate to Chevron.138 This is surely true if state agencies and legislatures would otherwise have inclined toward more capacious and flexible methods of statutory interpretation, but even, I think, if both are strict textualists.

Strikingly, Pappas concludes that the judicial review provisions of SLAPAs have little effect upon state deference standards; because they are silent as to this question, courts establish deference rules based upon their own jurisprudential preferences.139 Pappas sees in this a parallel to the federal situation, whereby Chevron is not required by nor based upon the federal APA.140 But this is a failure to recognize the sense in which Chevron accommodates, echoes, and matches, even though it does not rely upon, the federal statute. States seeking a similar congruence between their deference standards and their SLAPA will not all find Chevron as good a fit.

A further complication arises in those states that make direct constitutional provision for judicial review of agency action. Chevron itself is judge-made and legislation-influenced doctrine; the United States Constitution says little about agencies and nothing specific about judicial review. Indeed, fitting the modern regulatory state into a constitution silent on the subject has been the central preoccupation of federal administrative

135. COLO. REV. STAT. § 24-4-103(8)(a) (2013) (cited in Bell, supra note 19, at 838). 136. FLA. STAT. ANN. § 120.536(1) (West 2008) (cited in Bell, supra note 19, at 838–39). 137. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 843 n.9 (1984). 138. Bell, supra note 19, at 839 & n.201. 139. Pappas, supra note 19, at 1008. 140. Id.


law and one in which Chevron has played a substantial part. Some state constitutions, however, are not silent with respect to agencies or to judicial review. The Michigan Constitution, for example, specifically provides that ―[a]ll final decisions, findings, rulings and orders of any administrative officer or agency existing under the constitution or by law . . . shall be subject to direct review by the courts as provided by law.‖141 The Missouri Constitution is similar.142

The Michigan Supreme Court relies upon this constitutionalization of judicial review in rejecting a Chevron-style framework, holding that this passage assumes the general judicial power expressed elsewhere in the document.143 This argument is strong. The concepts of the judicial power and of judicial review that appear in and are derived from the United States Constitution assume a context very different from that of the regulatory state. But the state constitutions, many recently revised, should be read to assume a regulatory state. If those documents set out a right of judicial review, there is reason to think that they imagine courts reviewing agency interpretations of law de novo.

B. Agency Rules Review

Most states allow political officials to modify or rescind agency rules before they become effective.144 The most common device is to provide statutorily for legislative review, either by full legislatures or by committees or commissions within the legislatures.145 Less frequently, governors are given power to review rules.146 A few states have hybrid institutional arrangements.147 The states also vary widely in the powers, criteria, and timetables associated with rules review.148 These arrangements may be purely statutory or pursuant to constitutional authorization.149

141. MICH. CONST. art. 6, § 28. 142. MO. CONST. art. 5, § 18 (stating that agency action is subject to ―direct review by the courts as provided by law‖ and that ―such review shall include the determination whether the same [is] authorized by law‖); see also S.C. CONST. art. 1, § 22 (establishing a right of judicial review of agency action, without specifying the content of that right). 143. See In re Complaint of Rovas Against SBC Michigan, 754 N.W.2d 259, 265–67 (Mich. 2008). 144. Neal D. Woods, Political Influence on Agency Rule Making: Examining the Effects of Legislative and Gubernatorial Rule Review Powers, 36 ST. & LOC. GOV‘T REV. 174, 174 (2004). 145. Rossi, Institutional Design, supra note 102, at 1201. 146. See, e.g., infra notes 147, 169, 173–74 and accompanying text. 147. See, e.g., WYO. STAT. ANN. § 28-9-106 (2013) (providing for approval by a council which must then be ratified by the governor). 148. See Dennis O. Grady & Kathleen M. Simon, Political Restraints and Bureaucratic Discretion: The Case of State Government Rule Making, 30 POL. & POL‘Y 646, 649–54 (2002) (presenting a fifty-state survey documenting variation in practices of rules review); Robert W. Hahn, State and Federal Regulatory Reform: A Comparative Analysis, 29 J. LEGAL STUD. 873, 875–81 (2000); Woods, supra note 144, at 175–77 & tbls.1–2 (using data collected by Grady and Simon to develop a summary measure of political influence over state rulemaking, by state). 149. See, e.g., NEV. CONST. art. 3, § 1(2) (authorizing legislature to ―provide by law‖ for legislative review of agency action and the nullification of regulations by legislative action); N.J. CONST. art. 5, § 4, ¶ 6 (similar).


All rules review regimes—by empowering political branches to review, approve, modify or reject agency rules—clearly have implications for judicial review of such rules, and therefore for judicial deference. The provisions for legislative review in the 2010 MSAPA,150 for example, empower a legislative ―rules review committee‖ to review all newly adopted rules and all existing rules to determine whether the rule is a ―valid exercise of delegated legislative authority‖151 that is ―necessary to accomplish the apparent or expressed intent of the specific statute that the rule implements,‖152 and ―a reasonable implementation of the law.‖153 The parallels to judicial review are striking.

One obvious problem relating to legislative rules review is its possible unconstitutionality. Cognate arrangements at the federal level would be clearly invalid under INS v. Chadha.154 As Jim Rossi has documented, the Chadha analysis also has had, appropriately, significant purchase in many states.155 Indeed, Rossi argues that for constitutional reasons ―the future of legislative rules review—particularly where there is a veto and the state has no explicit constitutional authorization—is bleak.‖156 But, where constitutional challenges have yet to be brought or are unsuccessful, legislative review of rules matters institutionally when considering judicial deference. As a matter of doctrine, the empowerment of the legislature to approve, rescind, or modify rules reduces the plausibility of the Chevron assumption that legislatures intend to delegate the law-interpretation function to agencies. As a matter of politics, rules review changes the position of the judiciary in the separation of powers game that otherwise characterizes judicial review of rules.157

Neither of these conclusions is self-evident. In the federal context, whether an agency action gets Chevron deference, no deference, or deference of some other kind, the Congress always can revoke or modify any (prospective) agency action, by passing new legislation. That the Congress retains the power to override agency action with which it disagrees is among the justifications for Chevron‘s idea of deferring to agency interpretation of the statute. Should the legislature object to an agency interpretation, it can set it aside. With respect to legislative power, rules review does not seem terribly different from this arrangement, especially if rescission or modification pursuant to that review requires

150. The MSAPA drafters comment that the legislative rules review committee is a model ―widely followed in state administrative law.‖ REVISED MODEL STATE ADMIN. PROCEDURE

ACT § 702 cmt. (2010); accord Woods, supra note 144, at 175. 151. REVISED MODEL STATE ADMIN. PROCEDURE ACT § 702(b)(1) (2010). 152. Id. § 702(b)(3). 153. Id. § 702(b)(4). 154. 462 U.S. 919 (1983). 155. Rossi, Institutional Design, supra note 102, at 1203–16. 156. Id. at 1212; see also REVISED MODEL STATE ADMIN. PROCEDURE ACT § 703 cmt. (2010) (discussing ways to avoid Chadha problems in enacting systems for rules review). 157. See Lisa Schultz Bressman, Procedures As Politics in Administrative Law, 107 COLUM. L. REV. 1749, 1752 (2007); William N. Eskridge, Jr. & John Ferejohn, The Article I, Section 7 Game, 80 GEO. L.J. 523, 540–41 (1992).


bicameralism and presentment to the governor, which it often does.158 (To the extent that it does not, the constitutional deficiencies of the review statute become correspondingly more severe.159)

It has long been understood, however, that legislatures rarely legislate in order to repeal or revise interpretations delegated to agencies, even when agencies depart from a previous statute‘s intent, however defined.160 It is in light of this disability that legislatures rely upon administrative process; ―procedures allow politicians to prevent deviations before they occur.‖161 One feature of such procedure at the federal level is that agencies work on what McCubbins, Noll and Weingast call ―autopilot‖: legislators need not revisit, and indeed both hope and expect not to revisit, agency decisions.162 This is why Chadha matters.163 When states provide for legislative rules review, agencies therefore become more ―permeable‖ to political influence.164 Their rules are on the quotidian agenda of the sitting legislative review committee. That committee and its parent legislature are expected routinely to give effect to their own preferences with respect to agency action post facto, and they enjoy the benefit of existing procedures, political norms, bureaucratic routines, resources, staff, and expertise in doing so. The circumstances of a legislature that has created such procedures seem very different than the autopilot circumstances under which Chevron assumes that the legislature has delegated to an agency the task of construing ambiguity.165 The legislature expects to review and to approve not only the text going into the statute books but also the interpretations that come out. Such a legislature asks an interpreting agency to opine as to the meaning of the statute, but not to decide it authoritatively.

Legislation authorizing rulemaking, moreover, is a complex and multistage game. Some controversy has attached to the claim, advocated by McCubbins and his colleagues, that much of federal administrative procedure serves to entrench the policies of enacting coalitions.166 This argument suggests that federal law, by favoring procedures adopted by enacting coalitions and subjecting results under those procedures only to a legislative veto that is cumbersome, uncertain, and rare, favors the

158. See REVISED MODEL STATE ADMIN. PROCEDURE ACT § 703 & cmt. (2010); see also H. Harold Levinson, Legislative and Executive Veto of Rules of Administrative Agencies: Models and Alternatives, 24 WM. & MARY L. REV. 79, 82–84 (1982). 159. See Blank v. Dep‘t of Corr., 611 N.W.2d 530, 538–39 (Mich. 2000); REVISED

MODEL STATE ADMIN. PROCEDURE ACT § 703 cmt. (2010); Woods, supra note 144, at 174. 160. Matthew D. McCubbins, Roger G. Noll & Barry R. Weingast, Structure and Process, Politics and Policy: Administrative Arrangements and the Political Control of Agencies, 75 VA. L. REV. 431, 441 (1989). 161. Id. at 442. 162. See id. at 444. McCubbins and his colleagues mean something more specialized by ―autopilot,‖ but the general observation holds as well. 163. Sarah J. Poggione & Christopher Reenock, Political Insulation and Legislative Interventions: The Impact of Rule Review, 9 ST. POL. & POL‘Y Q. 456, 460 (2009). 164. Id. at 461, 466; accord Woods, supra note 144, at 174 (rules review gives state legislatures ―additional leverage‖ over agency outcomes). 165. See Poggione & Reenock, supra note 163, at 460–62. 166. For a review, see Steven J. Balla, Procedures and Political Control of the Bureaucracy, 92 AM. POL. SCI. REV. 663, 664 (1998).


preferences of enacting coalitions.167 But there can be no doubt that rules review deemphasizes the preferences of enacting coalitions in favor of sitting ones.168 This makes it even less plausible to assume, as Chevron does, that ambiguity is a delegation to agencies. A legislature that anticipates rules review does not anticipate that interpretative authority will be granted to an agency whose procedures work to implement the preferences of a previous legislature.

Furthermore, when legislatures review rules, courts engage in judicial review only of rules acceptable to those sitting coalitions (because unacceptable ones will be voided). Judges might therefore properly be inclined to be less deferential in determining whether a rule is consistent with a statute passed by an earlier legislature, in order to avoid giving sitting coalitions a too-powerful tool to impose their preferences without engaging in the ordinary process of legislation.

The other variety of rules review, less ubiquitous in the states than legislative review but still fairly common, is rules review by the governor. In a number of states, agency rules cannot become effective without gubernatorial approval, or may be rescinded by the governor, sometimes in conjunction with a commission charged with review of rules.169 This arrangement bears a strong family resemblance to the institution of federal regulatory review in the Office of Information and Regulatory Affairs (OIRA), part of the Executive Office of the President.

The resemblance is particularly strong when gubernatorial rules review is twinned with requirements that governors conduct cost-benefit or other economic analyses of rules, OIRA-inspired requirements that are becoming more popular in the states.170 There are also, of course, many salient institutional differences. For example, state statutes providing for gubernatorial review do not share the limitation that OIRA may review only ―significant‖ rulemakings.171 They provide for a clear veto of rules, rather than a process of ―return‖ and delay. Perhaps most important with respect to judicial review,172 gubernatorial review in some states is provided for by executive order, as is true for OIRA;173 in other states, however, it is provided for by statute.174

167. See McCubbins, Noll & Weingast, supra note 160, at 443–44. 168. This is true a fortiori in those states whose rules review procedures extend to already enacted rules. 169. See Woods & Baranowski, supra note 72, at 1220. 170. Cf. Edward Glaeser & Cass R. Sunstein, Moneyball for State Regulators, NAT‘L

AFFAIRS (forthcoming) (advocating state-level OIRA-type review). 171. See Exec. Order No. 12,866 § 6(b)(1), 58 Fed. Reg. 51,735 (Sept. 30, 1993); see also Exec. Order No. 13,563 § 1(b), 76 Fed. Reg. 3821 (Jan. 18, 2011) (reaffirming Exec. Order 12,866); Jennifer Nou, Agency Self-Insulation Under Presidential Review, 126 HARV. L. REV. 1755, 1768–69 (2013) (recounting history of this provision). 172. See Daniel A. Farber & Anne Joseph O‘Connell, The Lost World of Administrative Law, 92 TEX. L. REV. 1137, 1183 (2014) (―It is anomalous that [OIRA,] such an important feature of the regulatory state[,] has no statutory basis.‖). 173. See, e.g., Johnson, supra note 65, at 146 (describing Florida); L. Harold Levinson, Legislative and Executive Veto of Rules in Administrative Agencies: Models and Alternatives, 24 WM. & MARY L. REV. 79, 84–85 & n.25 (1982) (citing institutions for


Like OIRA, and also like legislative rules review, gubernatorial rules review raises various constitutional questions, which vary with particular arrangements. But again like OIRA (and also like legislative rules review), where the institution operates, it is relevant to establishing the appropriate deference framework for judicial review of rules. The debate over whether Chevron deference should apply to rules in which OIRA‘s handiwork is visible175 is roughly cognate to the issues raised by gubernatorial review in the states. On the one hand, rules review outside the agency might be thought to undermine the Chevron assumption that ambiguity is a delegation to the agency.176 On the other hand, gubernatorial review strengthens the applicability of Chevron‘s notion that deference is justified by the political power of the executive: ―While agencies are not directly accountable to the people, the Chief Executive is, and it is entirely appropriate for this political branch of the Government to make such policy choices‖177 necessary to resolve ambiguity, with appropriate regard to political considerations it views as relevant.178 In states where the statute empowers the governor to review rules, this latter argument may be somewhat stronger in the case of gubernatorial review than in that of OIRA; the statute creates quite clear political expectations about where the buck stops.

The issue of the nonunitary executive, however, complicates the issue further in the states. Gubernatorial rules review generally includes the power to review rules promulgated by agencies whose heads are independently elected. Like the institution of the state attorney general, therefore, such review creates an intraexecutive check on agencies‘ interpretive power. The implications of this for deference policy are muddled indeed. Because any rule promulgated by an agency with an independently elected head has also received gubernatorial approval, tacit or explicit, one might imagine that a high level of judicial deference is appropriate, because the agency‘s interpretation has already been vetted by another political actor. Judicial review, when it is the second review of agency action, might appropriately take particular pains to avoid deciding

gubernatorial review in Arizona, Connecticut, Hawaii, Nebraska, and Wyoming). The 1981 MSAPA offers states the option of instituting gubernatorial review whereby governors may ―rescind or suspend‖ any agency rule by executive order. See REVISED MODEL STATE ADMIN. PROCEDURE ACT § 3-202(a) (1981). The 1981 MSAPA also permits governors similarly to ―summarily terminate any pending rule-making proceeding by an executive order to that effect, stating therein the reasons for the action.‖ Id. § 3-202(b). There are no cognate provisions in the 2010 MSAPA, which does not discuss gubernatorial review at all, notwithstanding detailed provisions for legislative review. 174. See, e.g., HAW. REV. STAT. § 91-3(c) (West 2008); NEB. REV. STAT. ANN. § 84-908 (LexisNexis 2012). 175. See Farber & O‘Connell, supra note 172, at 1171–72; Shane, supra note 71, at 680. 176. See, e.g., Lisa Heinzerling, Statutory Interpretation in the Era of OIRA, 33 FORDHAM URB. L.J. 1097, 1116 (2006) (―[T]he theoretical reasons for Chevron deference do not apply in a situation where an expert agency like EPA takes its interpretive directions from OIRA.‖). 177. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 865 (1984). 178. See id.; Farber & O‘Connell, supra note 172, at 1172 (―[P]ortions of Chevron‘s reasoning seem to provide affirmative support for White House intervention.‖).


questions of ―policy.‖ On the other hand, as discussed below, gubernatorial review is like OIRA review and judicial review in creating a sequential game, where every rulemaking occurs in the shadow of the potential vetogates to come.179 Depending on the allocation of policy preferences and political needs within a plural executive, one can imagine situations in which gubernatorial review of rules promulgated by other elected executive officials could induce rulemaking outcomes less faithful to statutory intent than they would have been unencumbered by such review.180

IV. BUREAUCRATIC INSTITUTIONS

Whatever the constitutional theory of Chevron, and however complex its relationship with the APA, there can be no doubt that one of its major justifications is the comparative institutional advantage enjoyed by agencies when interpreting the statutes that they administer. Chevron urges deference in the situations where ―the force of the statutory policy in the given situation has depended upon more than ordinary knowledge respecting the matters subjected to agency regulations.‖181 One excellent reason to think that the Congress delegated to the EPA the task of interpreting the Clean Air Act is the latter‘s ―great expertise.‖182

Of course, Chevron accords deference to all agencies, not just ―expert‖ ones.183 This reliance upon a generalization—many agencies have expertise relevant to statutory interpretation, so we should defer to all agencies‘ statutory interpretation—is properly controversial. Sunstein argues that Chevron relies upon that generalization knowing it to be less than universal because simplicity is desirable given ―undeniable claims that specialized competence is often highly relevant.‖184 The suitability of Chevron to state administrative law depends, then, upon whether the generalization is generally accurate as to the states. If, as Sunstein argues it might be, state agencies‘ expertise is generally dwarfed by their unreliability, bias, incompetence, and low visibility,185 then Chevron is not a policy that matches the states.186

179. See Nicholas Bagley & Richard L. Revesz, Centralized Oversight of the Regulatory State, 106 COLUM. L. REV. 1260, 1270 (2006); Nina A. Mendelson & Jonathan B. Wiener, Responding to Agency Avoidance of OIRA, 37 HARV. J.L. & PUB. POL‘Y 447, 464 (2014); Nou, supra note 171, at 1757–58 (explicitly analogizing the strategic considerations created for agencies by presidential and judicial review). 180. A worked-out specification of this game would be of great interest but must be left to future work. 181. Chevron, 467 U.S. at 844. 182. Id. at 865. 183. See Shapiro & Fisher, supra note 8, at 476–81, 495 (delineating two different concepts of ―expertise,‖ one institutional and one deliberative and therefore contextual, and arguing that Chevron endorses the former). 184. Sunstein, supra note 11, at 2598. 185. Id. at 2596–97 & n.79. 186. Cf. Funk, supra note 54, at 172–73 (noting the argument that rationality review of state agency decisions might be inappropriate in light of state agency inexpertise, but concluding that ―[a]t the most, the difference in size, resources, and expertise between


Which is it? Are state agencies in general properly characterized as expert, or are they closer to Sunstein‘s ―parallel world‖ of incompetent and untrustworthy regulators?187 The latter suspicion is not without justification.188 Politicians and political scientists through the 1960s dismissed state government as a backwater at best. Jon Teaford catalogues representative thinking about the quality of state government in that period and finds this view to be the dominant one. He quotes journalist Robert Allen‘s characterization of state government as ―the tawdriest, most incompetent and most stultifying unit of the nation‘s political structure,‖ characterized by ―[v]enality, open domination and manipulation by vested interests.‖189 This sounds a great deal like Sunstein‘s alternate regulatory universe. Teaford then chronicles the rehabilitation of the states in the political science literature beginning in the 1970s;190 but academic lawyers and others continue to complain about the relative understaffing, underfunding, and low expertise of state as compared to federal agencies.191

Teaford, on the other hand, argues for himself that states were never the backwaters that academic and political elites thought them to be, nor were they utterly transformed in the late twentieth century. He characterizes the development of state agencies in the early twentieth century as the ―transfer of power from elected local officials to expert state bureaucracies.‖192 State agencies took part in the technocratic, expertise-loving Progressivism of the early twentieth century just as federal ones did, Teaford claims.193 It is certainly true that complex and sophisticated state agencies confront statutory regimes and complex policy problems in areas like transportation, environmental quality, K–12 schooling, and higher education. But even Teaford would agree, notwithstanding, that many other state agencies remain ―small, with little or no professional staff, and in some cases run on a part-time basis by persons whose primary jobs are elsewhere.‖194

Overall, lawyers, political scientists and others writing in this area still tend to give the state regulatory apparatus less sustained attention than the political branches. I therefore think it is fair to say that we do not know whether ―expert‖ is a fair characterization of state agencies in general. It is a question that awaits more work.

The place of state agencies in the federal system also vitiates the appropriateness of treating state agencies as expert in Chevron‘s sense of agency expertise. This is especially true of those state agencies that are

federal and state agencies . . . might call for some accommodation in the manner or ‗bite‘ of rationality review, but it does not require the abandonment of such review‖). 187. Sunstein, supra note 11, at 2596. 188. See Bruhl & Leib, supra note 48, at 1280; D. Zachary Hudson, A Case for Varying Interpretive Deference at the State Level, 119 YALE L.J. 373, 379 & n.24 (2009). 189. TEAFORD, supra note 78, at 2 (quoting ROBERT ALLEN, OUR SOVEREIGN STATE vii, xi–xii (1949)). 190. Id. at 4–5. 191. See, e.g., Funk, supra note 54, at 172. 192. TEAFORD, supra note 78, at 6. 193. Id. at 11–12. 194. Funk, supra note 54, at 168–69.


concerned with complex policy area and so in some respects are the most likely to be characterized by bureaucratic expertise of the kind relevant to Chevron. State officials who deal with the environment, education, or antiterrorism are enmeshed in a system of regulatory federalism195 that often very substantially deprives them of freedom of action. State and local policymakers, constrained by the legal supremacy of their federal counterparts and the sometimes asphyxiating regulatory requirements associated with fiscal federalism and the administration of federal spending programs, often cannot freely exercise their expertise with respect to policy.196 Their routine dependence on federal funds reduces their freedom of action.197 Moreover, they can find themselves in an adversarial relationship with federal agencies,198 and they are at a disadvantage in the fight. They therefore cannot be expert in the same way that federal agencies are.199 More specifically, in policy areas where regulatory responsibility is shared across federal and state levels and characterized by conflict, state agency statutory interpretation is likely to pay considerable attention to avoiding federal preemption, which poses a constant threat to state regulatory power.200 This environment of uncertain authority is one in which expertise in statutory interpretation is likely to be dominated by considerations of power and politics.

Regulatory federalism can also function without conflict; but so-called cooperative federalism201 vitiates agencies‘ expertise in its own way, and their political accountability along with it. State agencies may see themselves as clients or collaborators of their federal agency counterpart rather than part of a state administration accountable to the state political branches.202 Such agencies are inclined to defer to federal policymakers, who naturally seek consistency in their one-to-many relationship with the states.203 Such vertical, federal relationships may be more central to the agency‘s worldview and its fiscal and policy stability than its location in a hierarchy responsive to state political branches. An agency with this kind

195. See Miriam Seifter, States, Agencies, and Legitimacy, 67 VAND. L. REV. 443, 456 (2014). 196. See Gillian E. Metzger, Administrative Law As the New Federalism, 57 DUKE L.J. 2023, 2075–76 (2008). 197. Cf. Brudney & Hebert, supra note 73, at 200 (dependence on federal funds insulates state agencies from influence by legislatures, governors, and other external actors). 198. See DONALD F. KETTL, THE REGULATION OF AMERICAN FEDERALISM 12–14 (2d ed. 1987). 199. Seifter also notes the converse problem, that states, ―naturally driven by their home-state interests,‖ can undermine the ―expertise-based legitimacy‖ of federal administrative decisions. See Seifter, supra note 195, at 491. 200. See John Kincaid, From Cooperative to Coercive Federalism, 509 ANNALS AM. ACAD. POL. & SOC. SCI. 139, 144 (1990). 201. Id. at 140. 202. See Seifter, supra note 195, at 468, 475; James E. Skok, Federal Funds and State Legislatures: Executive-Legislative Conflict in State Government, in STATE POLITICS AND

THE NEW FEDERALISM: READINGS AND COMMENTARY (Marilyn Gittell ed., 1986). 203. See Nina A. Mendelson, Chevron and Preemption, 102 MICH. L. REV. 737, 779–80 (2004).


of multiple-master set of relationships across the federal system seems quite unlike the expert and linearly accountable agencies that Chevron posits.204

CONCLUSION: CHEVRON IN THE STATES

As I show in Part I, unlike many federal administrative law doctrines, Chevron has been received spottily by the states. Although a significant number of states have adopted deference doctrines that closely approximate Chevron‘s, few have endorsed it wholeheartedly. Only a few have rejected it explicitly either, but the majority of the states have deference doctrines meaningfully different from Chevron‘s.

In this Essay I have tried to show why this diversity is reasonable. Chevron is not generally a good fit for states. Institutional differences between the states and the federal government (and, for that matter, among the states), although they do not matter a great deal with respect to many aspects of administrative law, are central with respect to administrative deference to agency action. Chevron rests explicitly upon an institutional analysis. Its holding rests upon thick knowledge of how legislation proceeds in the Congress, what shapes the President‘s executive power, how federal courts exert their authority, and the scope and diversity of federal rulemaking and adjudication. All of these institutions are different in the states. State judges are chosen and retained differently.205 The judicial power itself is different.206 Executive power is different too, distributed across multiple executives who have complicated and overlapping contact with agencies.207 Legislation is different. The governing statutory regimes are different.208 And agencies themselves are different, arguably less expert and more ad hoc.209

If we take Chevron at its word that an institutional analysis justifies its holding, then there is no clear case for applying that holding in the states. At a minimum, Chevron with its institutional reasoning offers no particularly good reasons to think that state courts should construe ambiguity in agency-empowering state legislation as delegations to state agencies. Taken together, state/federal institutional differences suggest that Chevron does not offer a natural direction for state deference law. The case that has so dramatically and fundamentally reshaped the federal system is a poor candidate for policy diffusion.

204. This problem is multiplied in the case of regional single-purpose agencies, whose masters are the state (sometimes states), various local jurisdictions, and other stakeholders, although it might be more accurate to characterize such agencies as having these entities as constituents but having hardly any ―master‖ at all. These agencies are not always involved in rulemaking, but to the extent that they are, the proliferation of constituencies creates severe coordination problems, which may be an argument for deferential judicial review. See Liesbet Hooghe & Gary Marks, Unraveling the Central State, But How? Types of Multi-Level Governance, 97 AM. POL. SCI. REV. 233, 239–40 (2003). 205. See supra Part II.A. 206. See supra Part II.C. 207. See supra Part II.B. 208. See supra Part III. 209. See supra Part IV.


That Chevron‘s answer lacks an institutional justification in the states does not imply, however, that Chevron‘s question does not apply to the states. It surely does. The issue of judicial deference is vital in any modern regulatory state that authorizes both powerful agencies and judicial review. Every state court system has confronted this question. What should state courts do when they review an agency action that relies upon an agency‘s reasonable, but arguably incorrect, interpretation of its authorizing statute, particularly in light of the institutional factors that I have described here?

One thing it should certainly do is think institutionally. Just as Chevron offers a federal doctrine of deference that is constitutionally sound, jurisprudentially defensible, statutorily consistent, and encouraging of effective public administration, so too must state deference doctrine embody these values. Chevron is a model for states not by virtue of its holding but of its multi-institutional mode of analysis. Just as Chevron asks, state courts should ask as well: What do legislatures expect, or might they expect, when they draft ambiguous language? How closely tied are agencies to elected members of the executive branch? What actors in the system have what kinds of accountability? What is the nature of the political safeguards that prevent or discourage agency overreach? How much discretion will agencies generally have? How much discretion do they need in order to be effective?

It is a mistake for state courts to focus exclusively or primarily upon the judicial power, and the judicial duty to say what the law is, in thinking about the deference problem.210 As too few state courts seem to realize, judicial deference is not a matter of judges declining to say what the law is; it is a matter of how judges decide what the law is.211 And the answer to that question must be responsive to the full range of a state‘s constitutional institutions, to the design of its administrative code, and to the nature and missions of its bureaucracies.

This Essay also suggests something about the content of state court deference rules. One straightforward conclusion is that states need not reject Chevron.212 The particular way in which Chevron balances between politics, the rule of law, and bureaucratic effectiveness is a worthy model for states to consider. Especially in those states whose institutional structures more closely resemble their federal counterparts—appointed, life-tenured judges, relatively unitary executives, expert agencies—Chevron could well be a good fit. And indeed many states have adopted Chevron or something like it. Nothing I have said suggests that they are making a mistake. It is not necessary for every aspect of Chevron‘s analysis to match state circumstances for it to be a useful model.213

210. See Sunstein, supra note 11, at 2610 (―The meaning of statutory enactments is no brooding omnipresence in the sky.‖). 211. But see Beermann, supra note 9, at 788–90. 212. Contra Hudson, supra note 188, at 380–81. 213. As I have noted, because Chevron is multivalent, it does not lose its appeal if, as many have argued, it is deficient in aspects of its analysis as applied to federal institutions.


A broader question is whether states should strive to set a deference rule at all. Are institutional diversity and nuance reasons to tailor deference rules to each state‘s institutional circumstances, but then to apply the tailored rule within that state across the board? Or should they be dominant considerations intrastate as well as across the states, so that state courts would tailor deference to the particular form of agency action at issue, the particular agency, the particular statute, or even the particular question presented? If the latter, then states would adopt no deference rules, establishing at most only deference ―canons,‖ methodological predispositions that courts could use or disregard based upon their sense of each situation, and apply or not without fear of creating or violating methodological precedent with respect to deference in future cases.214

As other contributors to this symposium make clear, this possibility is alive and well at the federal level, where scholars are pressing claims that Chevron in fact operates less as a rule in federal administrative law than as a canon irregularly applied—and that this is normatively appropriate or even desirable. On this view, there should be no more ―methodological stare decisis‖ when it comes to judicial review of agency statutory interpretation than for any other form of statutory interpretation.215 Instead federal courts should ―tailor deference to variety.‖216

Demurring to the merits of these positive and normative claims at the federal level, which are far outside the scope of this Essay, one might ask whether the institutional differences among states and between states‘ systems and the federal system ought to affect the outcome of these questions. Is there anything about particularly state constitutional institutions, state administrative-law statutes, or state bureaucracies that urge greater methodological stare decisis, or greater ad-hocness, than would be appropriate at the federal level?

I would point to two. One is at the level of state courts, which are characterized by a higher level of methodological stare decisis with respect to interpretation than federal courts.217 The other is a point already made about state bureaucracies: their expertise and quality is more variable than federal agencies‘. If state agency action is already particularly prone to ad hoc and unsupportable decision-making, then clear rules of the road mitigate these problems, even at the level of generality associated with a deference rule.218 If an agency knows that it is entitled to deference of a

214. See Gluck, supra note 53, at 1761. 215. See Abbe R. Gluck, What 30 Years of Chevron Teach Us About the Rest of Statutory Interpretation, 83 FORDHAM L. REV. 607, 612–13 (2014); accord Shane & Walker, supra note *, at 491. 216. United States v. Mead Corp., 533 U.S. 218, 236 (2001). 217. See Gluck, supra note 53, at 1754. 218. Clear deference rules can also be expected to mitigate statutory ambiguity at the state level. See Robert J. McGrath, Legislatures, Courts, and Statutory Control of the Bureaucracy Across the U.S. States, 13 ST. POL. & POL‘Y Q. 373, 374 (2013).


particular kind (including no deference at all), it actions are less ad hoc than when it is guessing as to what courts might do.219

Of course, under a rule that courts review agency statutory interpretation de novo, agencies are already guessing; but at least they need not guess as to whether they should be guessing. Anticipating de novo review, agencies can seek to bulletproof their claims and respond to all arguments; anticipating deference, they can seek to show reasonableness and emphasize their freedom to act under ambiguity. But when deference is ―tailor[ed] to variety,‖220 the predictability of administrative action is less certain; and predictability is more badly needed at the state than at the federal level.

Ultimately, however, the critical lesson is that states should not miss how involved, and how important, the deference question is. Tailoring deference to variety cannot be allowed to slide into fudging the complexity of the deference issue. State courts must think about their relevant institutions—all of them—when they decide whether to defer to agency judgment about what statutes mean. In signaling the importance of that undertaking, Chevron offers a beacon even to those state courts that will, properly, choose to reject its holding.

219. Cf. Scalia, supra note 9, at 517 (―The legislative process becomes less of a sporting event when those supporting and opposing a particular disposition do not have to gamble upon whether, if they say nothing about it in the statute, the ultimate answer will be provided by the courts or rather by the Department of Labor.‖). A parallel argument can be made about state legislatures‘ grants of powers to agencies under clear deference rules. 220. Mead, 533 U.S. at 236.

Documents

CHEVRON AND DEFERENCE IN STATE ADMINISTRATIVE LAWfordhamlawreview.org/wp-content/uploads/assets/pdfs/Vol_83/Saiger... · 555 CHEVRON AND DEFERENCE IN STATE ADMINISTRATIVE LAW Aaron